Abstract:Background and Objective: Diseases and pests are one of the major reasons for low productivity of apples which in turn results in huge economic loss to the apple industry every year. Early detection of apple diseases can help in controlling the spread of infections and ensure better productivity. However, early diagnosis and identification of diseases is challenging due to many factors like, presence of multiple symptoms on same leaf, non-homogeneous background, differences in leaf colour due to age of infected cells, varying disease spot sizes etc. This paper is an attempt to overcome these challenges and provide apple disease detection and identification system that can perform real-time detection of these diseases with higher accuracy and faster detection speed.Methods: In this study, we first constructed an expert-annotated apple disease dataset of suitable size consisting around 9000 high quality RGB images covering all the main foliar diseases and symptoms. Next, we propose a deep learning based apple disease detection system which can efficiently and accurately identify the symptoms. The proposed system works in two stages, first stage is a tailor-made light weight classification model which classifies the input images into diseased, healthy or damaged categories and the second stage (detection stage) processing starts only if any disease is detected in first stage. Detection stage performs the actual detection and localization of each symptom from diseased leaf images. Both classification and detection models were initialized using transfer learning and then trained on prepared dataset. Data augmentation techniques like rotation, translation, noise, cut-out etc. were also applied to prevent overfitting. Results: The proposed approach obtained encouraging results, reaching around 88% of classification accuracy and our best detection model achieved mAP of 42%. The qualitative results validate that the proposed system is effective in detecting various types of apple diseases and can be used as a practical tool by farmers. The preliminary results of this study look promising even on small or tiny spots.Conclusion: The proposed apple disease detection system achieved promising results and can be very helpful tool for farmers and apple growers to aid them in diagnosis, quantification and follow-up of infections. Furthermore, in future, the work can be extended to other fruits and vegetables as well.

SeptoSympto: a precise image analysis of Septoria tritici blotch disease symptoms using deep learning methods on scanned images

Towards high throughput in-field detection and quantification of wheat foliar diseases with deep learning

Deep Learning-Based Barley Disease Quantification for Sustainable Crop Production

SYMPATHIQUE: image-based tracking of symptoms and monitoring of pathogenesis to decompose quantitative disease resistance in the field

Evaluation of Rice Bacterial Blight Severity from Lab to Field with Hyperspectral Imaging Technique.

Laboratory and UAV-Based Identification and Classification of Tomato Yellow Leaf Curl, Bacterial Spot, and Target Spot Diseases in Tomato Utilizing Hyperspectral Imaging and Machine Learning

Digital image quantification of rice sheath blight: Optimized segmentation and automatic classification

ScAnalyzer: an image processing tool to monitor plant disease symptoms and pathogen spread in Arabidopsis thaliana leaves

Deep Diagnosis: A Real-Time Apple Leaf Disease Detection System Based on Deep Learning

Autonomous Detection of Plant Disease Symptoms Directly from Aerial Imagery

Image-Based Wheat Fungi Diseases Identification by Deep Learning

Early detection and spectral signature identification of tomato fungal diseases (Alternaria alternata, Alternaria solani, Botrytis cinerea, and Fusarium oxysporum) by RGB and hyperspectral image analysis and machine learning

Deep Learning for Tomato Diseases: Classification and Symptoms Visualization

Quantifying Leaf Symptoms of Sorghum Charcoal Rot in Images of Field-Grown Plants Using Deep Neural Networks

Image Quality Assessment for Foliar Disease Identification (AgroPath)

Colour-analyzer: a new dual colour model-based imaging tool to quantify plant disease

Deep semantic segmentation for the quantification of grape foliar diseases in the vineyard

Image set for deep learning: field images of maize annotated with disease symptoms

Optimizing Corn Tar Spot Measurement: A Deep Learning Approach Using Red-Green-Blue Imaging and Stromata Contour Detection Algorithm for Leaf-Level Disease Severity Analysis

Enhancing practicality of deep learning for crop disease identification under field conditions: insights from model evaluation and crop-specific approaches

Fusarium head blight detection, spikelet estimation, and severity assessment in wheat using 3D convolutional neural networks